Automatic detection and classification of liver cancer based on neural network techniques

Abstract

Cancer is the leading cause of mortality in every region of the globe. newlineThe early detection and diagnosis of liver cancer is one of the most vitally newlineimportant parts of cancer prevention. In clinical radiology, two of the difficult newlineproblems to solve in liver cancer are finding it early and correctly classifying newlineit. The objective is to detect liver cancer at an early stage in order to save as newlinemany lives as possible. newlineThe objective of this research is to develop an automated method that newlinecan reliably identify and categorize liver cancers utilizing an image processing newlinesystem. This method will be of use to the radiologist in automatically newlinerecognizing any lesions that have occurred in the liver. Because, manually newlinedetecting cancer tissue is a tedious and time-consuming activity. Computer newlineAided Diagnosis (CAD) has been greatly employed to assess and analyze liver newlinecancer in order to accurately and rapidly diagnose the disease. The Computed newlineTomography (CT) scan has emerged as an essential diagnostic tool for liver newlinecancer. newlineThe primary motive of this research is to develop a method that may newlinebe used for classifying and dividing liver cancer into several subtypes. The liver newlinewas autonomously segmented from the full abdominal CT scan image by using newlineactive contour, level set and PSO algorithms. The features were extracted from newlinethe liver that had been segmented once the segmentation process was complete. newlineThese generated features were used in the classification of liver cancer by the newlineapplication of techniques such as Random Forest (RF), Feed Forward Neural newlineNetwork (FFNN) and Back Propagation Neural Network (BPNN). newline